import torch import torch.nn.functional as F import numpy as np def print_memory(device, tag=""): if torch.cuda.is_available(): print(f"{tag} | 当前显存: {torch.cuda.memory_allocated(device) / 1024**2:.2f} MB, 最大显存: {torch.cuda.max_memory_allocated(device) / 1024**2:.2f} MB") else: print("当前设备不支持 cuda.") def nms_sampling(noise: np.ndarray, k: int, radius=2): # noise: [H, W] noise = noise.copy() points = [] for _ in range(k): idx = np.argmax(noise) x, y = np.unravel_index(idx, noise.shape) points.append((x, y)) # 抑制周围 x0 = max(0, x - radius) x1 = min(noise.shape[0], x + radius + 1) y0 = max(0, y - radius) y1 = min(noise.shape[1], y + radius + 1) noise[x0:x1, y0:y1] = -np.inf result = np.zeros_like(noise) for x, y in points: result[y, x] = 1 return result def masked_focal( logits: torch.Tensor, target: torch.Tensor, tile_set: set, gamma: float = 2.0, ) -> torch.Tensor: """ 通道专属 Focal Loss:tile_set 内的位置以真实 tile ID 为目标, tile_set 外的位置以 0(空地)为目标,全部位置均参与损失计算。 这样模型不仅要学会"这里是什么 tile",还要学会"这里不应该是本通道的 tile", 避免解码器在所有位置都输出专属类别来规避损失。 Args: logits: [B, H*W, num_classes] 解码头输出(未经 softmax) target: [B, H*W] 完整地图 ground truth(整数 tile ID) tile_set: set of int 本通道专属 tile 集合 gamma: Focal Loss 聚焦参数 Returns: scalar tensor 通道专属 Focal Loss(均值) """ B, S, C = logits.shape # 非专属 tile 位置目标替换为 0(空地),专属 tile 位置保持原始标签 in_set = torch.zeros(B, S, dtype=torch.bool, device=logits.device) for t in tile_set: in_set |= (target == t) corrected = target.clone() corrected[~in_set] = 0 # Focal Loss,全部位置参与计算 ce = F.cross_entropy( logits.view(-1, C), corrected.view(-1), reduction='none', ).view(B, S) # [B, S] pt = torch.exp(-ce.detach()) # 正确类预测概率,stop-gradient fl = (1.0 - pt) ** gamma * ce return fl.mean()